Proxy panels enable privacy-aware outsourcing of genotype imputation

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genomic data sharing is crucial for research but poses privacy challenges, especially when outsourced for analysis like genotype imputation.
• Existing cryptographic methods for privacy are often computationally intensive, require specialized expertise, and are not scalable for population-level studies.

Purpose of the Study:

• To develop practical privacy-preserving protocols for outsourcing genotype imputation using reference panels.
• To protect sensitive genotype panel data while enabling external analysis.

Main Methods:

• Developed protocols using "proxy panels" generated through haplotype sampling, allele hashing, and coordinate anonymization.
• Proxy panels are designed to be distinct from sensitive panels but usable as input for imputation software (e.g., Beagle).
• Evaluated the effectiveness of proxy-based imputation against known security attacks and its impact on imputation accuracy.

Main Results:

• Proxy panels successfully protect sensitive genetic variant coordinates, genetic maps, and haplotypes.
• Proxy-based imputation demonstrates protection against common security threats.
• A minor decrease in imputation accuracy was observed across various allele frequencies.

Conclusions:

• Proxy panels offer a viable solution for privacy-preserving outsourcing of genotype imputation.
• This approach balances data security with the utility of genomic data for large-scale collaborative research.
• The method is practical for population-scale genomic analyses, addressing limitations of current cryptographic techniques.